Learned Control Variates for Efficient EV Charging Grid-Outage Estimation
October 21, 2026
To reduce the computational cost of Monte Carlo simulations for rare grid outages, this method uses a learned classifier trained on observable features like temperature and load. The variance reduction achieved is an explicit function of the classifier's recall and false-positive rate.
HOW THIS AFFECTS YOU
●
researcherYou can use cheap, observable feature classifiers as control variates to accelerate high-fidelity grid simulations.